Ask Smithsonian: Can Animals Predict Earthquakes?

Humans cling to the idea that the beasts of the field could help to make earthquakes predictable, but prognosticators they are not

Earthquakes are frightening events, striking without notice. But some believe there is an early-warning system: animals.

Over the centuries, people have reported seeing animals head for the hills or leave their burrows in the weeks, days or hours before a temblor hits. But is this belief grounded in science?

It’s true that animals can sense a quake, usually just minutes before humans do, says Michael Blanpied, associate coordinator of the U.S. Geological Survey Earthquake Hazards Program. Established by Congress in 1977, the program monitors and reports earthquakes, assesses earthquake impacts and hazards, and researches the causes and effects of earthquakes.

But that’s a reaction, Blanpied adds, not a special talent for predicting when or where a quake might hit. 

Some researchers have theorized that certain creatures can detect signals that humans cannot, such as subtle tilting of the ground, changes in groundwater or variations to electrical or magnetic fields.

Seismologists would love to have an earthquake early-warning system, but animals don’t appear to be the answer, says Blanpied.

“The most likely time to have a big earthquake is after a small quake,” he says. But even knowing that little quakes beget big ones isn’t much help. One quake doesn’t give scientists the ability to know exactly how long until the next one, or even where it will have its epicenter. Tracking quakes is no easy feat given that the Earth experiences millions a year, many of which are barely noticed.

Humans have a longstanding attachment to the belief that animals know a quake is on the way. Throughout history—starting in ancient Greece—animals have reportedly been observed fleeing an area that subsequently had an earthquake. The observations, however, were recalled in hindsight, after the quake.

It’s hard to document—before a quake happens—that animal behavior changed, especially since quakes happen without warning, says Blanpied. The USGS sponsored a project in the late 1970s to continuously observe lab rodents in southern California to see if there was a burst of activity just before a quake. Unfortunately, there were no quakes during the study’s duration.

Jim Berkland, a San Francisco Bay Area geologist, made a name for himself by accurately predicting Northern California’s 1989 Loma Prieta earthquake. His forecast relied, in part, on combing the classified ads for local newspapers, which he said demonstrated that a larger than usual number of household pets were listed as missing in the week or so before the 6.9 magnitude quake.

Berkland was not the only one to claim that missing pets indicated something afoot. The USGS says, however, that the missing pet theory does not hold water, and as proof points to a 1988 study disputing the claim.

The agency does not outright dismiss the possibility of animal activity as a predictor, says Blanpied. It’s just that the USGS has not received many funding requests for such studies, and is not likely to conduct them on its own, he says. On its website, the agency points to a 2000 study by seismologist Joseph L. Kirschvink, which suggested that animals’ instinctual fight or flight response may have evolved over the millennia to also be a sort of early warning system for seismic events. He suggested ways to study how animals might react to potential precursors of earthquakes, such as ground tilting, groundwater changes or electrical or magnetic field variations.

Many who believe animals can sense quakes point to work done by Friedemann T. Freund, who is a senior research scientist at the non-profit SETI Institute (which is searching for extraterrestrial life). He has postulated for decades that rapid stresses in the earth’s crust just before a quake cause major changes in magnetic fields, which animals can sense. Blanpied says these theories “have been roundly questioned and criticized,” because rapid stress changes would not be expected before a quake, and because such changes were never observed or recorded outside of Freund’s lab.

Freund remains undaunted. In 2015, he and co-researchers published a study showing that animals in Peru’s Yanachaga National Park basically disappeared in the weeks leading up to a 7.0 magnitude quake in the region in 2011.

Animals are able to detect the first of an earthquake’s seismic waves—the P-wave, or pressure wave, that arrives in advance of the S-wave, or secondary, shaking wave. This likely explains why animals have been seen snapping to attention, acting confused or running right before the ground starts to shake, Blanpied says. Also, some animals—like elephants—can perceive low-frequency sound waves and vibrations from foreshocks that humans can’t detect at all.

Just ahead of the 5.8 magnitude quake that hit the Washington, D.C. area in 2011, some of the animals at the Smithsonian Institution’s National Zoo raised a ruckus, says Kenton Kerns, a small mammal biologist at the Zoo. Among those were the lemurs, who began “calling”—loudly vocalizing—about 15 minutes before keepers felt the ground shaking. Keepers recalled the activity after the quake took place.

Lemurs—small primates from Madagascar—are prolific vocalizers when upset, and can make their grievances known multiple times a day, says Kerns. Which means it’s not possible to know whether they sensed the impending quake, or if something else coincidentally perturbed them, he says.

So why do humans cling to the idea that animals are prognosticators? “I think people feel comforted by the idea that there would be something that would make earthquakes predictable,” says Blanpied.

UPDATE 8/11/2016: Due to an editing error, a previous version of this article implied that Dr. Freund had passed away. We regret the error. 

Get the latest on what's happening At the Smithsonian in your inbox.